US pharmaceutical pricing runs on two ledgers: a public list price and a confidential net price, with the gap between them negotiated in private. German pricing for new medicines has run, since the 2011 AMNOG reform, on one ledger. The reimbursement amount negotiated with the statutory health insurers after benefit assessment is published, and pricing authorities far beyond Germany read it.1Arzneimittelmarktneuordnungsgesetz (AMNOG) vom 22. Dezember 2010 (BGBl. I S. 2262), in force 1 January 2011. The Medizinforschungsgesetz (MFG) of 23 October 2024 rebuilt part of that architecture with effect from 1 January 2025, and it attached both of its new pricing levers to the same variable: whether the company's research happens in Germany.2Medizinforschungsgesetz (MFG) vom 23. Oktober 2024 (BGBl. 2024 I Nr. 324); § 130b amendments in force since 1 January 2025. A trial enrollment map that was fixed years before launch, and the location of a research department, have become pricing terms in the largest pharmaceutical market of the EU. Most US-headquartered companies encounter the connection after both were decided.
1. The Confidential-Price Mechanism and Its Research Condition
The default sequence is familiar to anyone who has launched in Germany. A medicine with a new active ingredient enters the market at a price the company sets. The G-BA, the self-governance body that runs the benefit assessment, scores the additional benefit against a comparator therapy, and the company then negotiates the Erstattungsbetrag, the reimbursement amount, with the GKV-Spitzenverband, the umbrella organization of the statutory health insurance funds; since the 2022 financial-stabilization reform the negotiated amount reaches back to the seventh month after launch.3GKV-Finanzstabilisierungsgesetz vom 7. November 2022 (BGBl. I S. 1990): negotiation guardrails; Erstattungsbetrag applies from the seventh month. The result is published. It flows into the pricing directories that pharmacies, wholesalers, and payers bill against, and it is the number the rest of the world sees.
The MFG adds a decision point five days wide. Within five days of an Erstattungsbetrag agreement concluded by 30 June 2028, or of an arbitral determination of the Schiedsstelle within the same window, the company may declare to the GKV-Spitzenverband that the amount will stay out of the public price-reporting chain; the election exists only in connection with the first placing on the market of a medicinal product with a new active ingredient.4SGB V, § 130b Abs. 1c: election, five-day window, research conditions, 9 percent Abschlag, confidentiality until expiry of Unterlagenschutz. The condition is evidentiary and territorial: documents demonstrating a drug-research department of the company in Germany (an Arzneimittelforschungsabteilung) and, in addition, relevant own projects and cooperations with public institutions in preclinical or clinical drug research in Germany. The GKV-Spitzenverband has seven days to determine whether the evidence suffices; if it finds against the company, the Schiedsstelle redecides within seven further days. What confidentiality buys is defined as precisely as what it costs: the reimbursement amount becomes the agreed figure minus a discount of 9 percent, and the substitution in the reporting chain holds until regulatory data protection for the product, its Unterlagenschutz, falls away.
Confidential does not mean secret. The price the market sees remains the company's own launch price; the funds settle pharmacy claims on that visible basis and recover the difference from the company afterwards through a statutory settlement mechanism. Hospitals, the authorities that supervise them, drug importers, and any legal person that can show it bought the product can obtain the actual amount on request, applicants for generic authorization can obtain it from twelve months before data protection expires, and the company bears a fee for each disclosure. The first election under the mechanism was reported in July 2025, for tirzepatide (Mounjaro); as of the Kassenärztliche Bundesvereinigung's October 2025 account, it remained the only one.5KV Berlin, Praxis-News (31 July 2025); KBV, Praxisnachrichten (2 October 2025): tirzepatide as the first and, to that date, only confidential reimbursement amount.
Confidentiality in the German system is not a privacy setting but a purchased position: it costs nine percent of the negotiated amount, the election window is five days, and only companies that can prove their research lives in Germany may buy it.
2. How the MFG Overlays the AMNOG Value Assessment
The second lever sits deeper in the negotiation mechanics. The 2022 GKV-Finanzstabilisierungsgesetz drew hard statutory boundaries, the so-called Leitplanken, around what may be agreed for products the G-BA scores modestly: a product without additional benefit that cannot be placed in a fixed-reference-price group must land, against a comparator that itself still enjoys patent or data protection, at least 10 percent below that comparator's annual therapy costs, and a product with minor or non-quantifiable additional benefit may not exceed the costs of such a comparator at all. For a US reader the nearest anchor is the absence of any such trade at home: the Inflation Reduction Act publishes the negotiated Medicare maximum fair prices, and no US federal pricing program offers relief from a pricing constraint in exchange for domestic trial enrollment. German law does exactly that. Where the G-BA states in its benefit-assessment decision that the product's clinical trials were conducted to a relevant extent in Germany, the two strictest guardrails cease to apply: the minus-10-percent rule gives way to a softer ceiling at the comparator's level, and the hard cap on minor and non-quantifiable benefit disappears.4SGB V (n 4), § 130b Abs. 3 Satz 11 bis 15: guardrail exception, three-year termination, thirty-month research-footprint proof.
The statute says what a relevant extent is: the share is relevant where at least five percent of the participants in the product's clinical trials, conducted or commissioned by the company, were enrolled at trial sites in Germany, for products first placed on the market from 1 January 2025. What the G-BA supplies is the machinery rather than the threshold, restating the test in its procedural rules and specifying the dossier evidence through which the enrollment share must be shown, and stating the resulting determination in the benefit-assessment decision itself.6Statutory five-percent test in § 35a SGB V; G-BA, Beschluss zur Änderung der Verfahrensordnung, 5. Kapitel (5 December 2024, in force 20 March 2025): dossier evidence. The relief is also conditional in time. For products that used the exception, the GKV-Spitzenverband must terminate the price agreement after three years unless the company demonstrates, thirty months in, the same research-department-and-cooperations footprint that gates the confidentiality election; a company that fails the proof renegotiates without the exception. Enrollment geography opens the door, and only institutional research presence keeps it open. A company that ran its pivotal trials partly in Germany but then moved its research elsewhere holds the benefit for three years and loses it prospectively.
3. What Counts as a German Research Footprint
The two levers rest on two different footprint tests, and the differences are where the difficulty lives. The trial-share test is arithmetical and retrospective: participants at German sites divided by total enrollment across the product's relevant trials, assessed by the G-BA on the dossier, for an indication whose studies were designed and enrolled years earlier. The institutional test is qualitative and immediate: a research department "of the company" in Germany, plus relevant own projects, plus cooperations with public institutions in preclinical or clinical drug research, assessed by the GKV-Spitzenverband within seven days and, on escalation, by the Schiedsstelle within seven more. One is decided by a body applying published procedural rules; the other by the counterparty to the price negotiation, on undefined statutory terms, in a fourteen-day corridor. The corridor is not the only route to an answer: a company that files the footprint evidence within six months of first placing the product on the market obtains the same seven-day determination, with the same escalation to the Schiedsstelle, in advance and valid for a year. What the six-month route buys is timing rather than certainty, since the substantive terms it is judged against are the same undefined ones.
Each element of the institutional test opens a question the statute does not answer. The pharmazeutische Unternehmer that signs the Erstattungsbetrag agreement is, in a US-headquartered group, typically a German distribution subsidiary; whether a research department of an affiliate, a parent, or an acquired German biotech is a department "of the company" is exactly the kind of attribution question group structures were not built to answer. What volume or quality makes own projects "relevant" is unstated. Cooperations qualify when they are with public institutions, which places university hospitals on one side of the line and raises the question where contract research, private-clinic networks, and industry consortia fall. The five-percent arithmetic has its own fine print: which studies enter the denominator, how multi-indication programs are counted, and what an ongoing global trial contributes are questions answered, if at all, in procedural rules a US clinical-operations team has usually never read.
4. Relocation, Partnering, and the Price-Confidentiality Trade
Whether the trade is worth taking is not a German question. The published Erstattungsbetrag travels: European pricing authorities reference it directly or indirectly, and Germany sits in the nine-country basket that anchors Switzerland's Auslandpreisvergleich, the external price comparison examined in Insight 23. A confidential amount leaves the visible German datum at the company's own launch price. Nine percent of German revenue is therefore the premium for shielding every downstream market that keys off the German number, and the value of that shield depends on how much revenue sits behind the reference point, product by product. The calculation acquired a US dimension within months of the mechanism going live: Executive Order 14297 of 12 May 2025 directs the Secretary of Health and Human Services to communicate most-favored-nation price targets benchmarked against prices in comparably developed nations, and manufacturer-by-manufacturer pricing agreements followed from autumn 2025.7Executive Order 14297 of May 12, 2025, 90 Fed. Reg. 20749: most-favored-nation price targets benchmarked against comparably developed nations; first manufacturer agreements announced 30 September and 10 October 2025. How a most-favored-nation benchmark treats a country whose net price is deliberately dark is a question neither legal order answers, and a mechanism built against European reference cascades faces an American audience it was not designed for.
The corporate-structure questions are just as concrete. The entity that holds the election is the German counterparty to the price agreement, while the research that must be proven usually sits elsewhere in the group; a footprint assembled through an affiliate, a university partnership, or an acquired discovery unit may or may not read as a department "of the company" on the day the GKV-Spitzenverband opens the file. In-licensed products sharpen the problem: whose research department, and whose projects, count when the German marketer did not originate the molecule? A divestment or a global research consolidation executed for reasons that have nothing to do with Germany intersects the thirty-month proof and the three-year termination, converting a portfolio decision into a pricing event. And the five-day election window means the confidentiality decision cannot be made when it arises; it must be made in advance, inside the pricing governance that also weighs US reference exposure, investor communication, and the physician-facing consequence that German prescribers cannot see the price at all, a consequence the medical profession and the funds were still working into the statutory efficiency-review machinery in late 2025.
5. Strategic Considerations
The deeper difficulty is temporal. Enrollment shares for products that will reach German pricing in 2028 or 2029 are being fixed in protocols approved before anyone knows whether the incentives will still exist. The confidentiality election is open only for agreements concluded by 30 June 2028, and the MFG obliges the federal health ministry to report on the effects of its § 130b changes by 31 December 2026. Should a trial footprint be priced as an option on a regime with a published expiry date, and if so, at what premium per German site? A company that treats the five-percent share as a pricing asset must also decide who inside the organization owns it: clinical operations chooses countries on recruitment speed and cost, market access inherits the consequences, and the two rarely share a planning horizon.
The legislative ground was moving as this article went to press. A Referentenentwurf of 16 April 2026 for a GKV-Beitragssatzstabilisierungsgesetz proposed striking the AMNOG guardrails altogether; the GKV-Spitzenverband rejected the guardrail deletion emphatically in its consultation response while welcoming the loss of the German trial-share machinery that goes with it, on the ground that rewarding domestic research out of health-insurance contributions is a task for the tax base rather than the funds. The Bundeskabinett adopted the government bill on 29 April 2026, and the Bundestag gave it a first reading on 12 June 2026 before referring it to committee.8GKV-Beitragssatzstabilisierungsgesetz, Referentenentwurf vom 16. April 2026, Regierungsentwurf vom 29. April 2026, BT-Drs. 21/6130; GKV-Spitzenverband, Stellungnahme vom 19. April 2026; proposed § 130a Abs. 1c SGB V. If the guardrails fall for every product, the § 130b exception loses its object, but the reward for German enrollment does not fall with it. The same bill re-homes the five-percent test in a new § 130a SGB V exemption from the proposed manufacturer discount, for products first placed on the market from 1 January 2027 and decided by the Bundesinstitut für Arzneimittel und Medizinprodukte rather than the G-BA. The incentive migrates rather than disappears, but it does not migrate unchanged: the government bill grants the exemption only where the trial share is met and, cumulatively, where production of the active substance in Germany is expected to contribute relevantly to needs-based supply. Same arithmetic, a second condition bolted onto it, a different statutory home, a different deciding body, and a start date that lands after the 2025 site selections that were meant to earn it. Whether the location premium a company built into those selections survives to its 2027 launch depends on a bill still in passage as of publication, which is an uncomfortable basis for a nine-figure trial budget.
The governance questions compound from there. What evidence file will carry the thirty-month proof, and who maintains it across reorganizations? May a US parent disclose, in a securities filing or an investor call, the number its German subsidiary elected to keep out of the price lists, and what do the transfer-pricing file and the auditors do with a price that is confidential in one direction only? What happens at the end is also part of the beginning: when Unterlagenschutz expires the substitution ends, generic applicants will have seen the number a year earlier, and the company re-enters price transparency at whatever level years of confidential settlement produced. None of these questions resolves in the abstract. They resolve against a specific portfolio, a specific comparator landscape, a specific enrollment map, and a specific tolerance for holding a price position whose legal architecture is scheduled for review, and that is where the analysis has to begin.